One horse. One bet. Every day. Sly Man’s read on the day’s card — the shortlist, the price, and the case for the pick.
Browse the logbookWhen the money talks this loud on a maiden filly, you stop and listen. Today's scan covered the declared card for Saturday 18 September 2026, running 551 runners through a four-stage filter pipeline before scoring what survived. The filters are: odds band 2.0–50.0 decimal (no odds-on, nothing beyond 49/1); field size 6–14 runners (trimming out chaotic big fields and statistically noisy small ones); and a lay-of-day collision check that dropped 1 runner flagged by the opposing book. Feature scoring runs seven dimensions — speed, form, class change, trainer strike rate, going, market move, and draw — weighted at 1.5, 1.2, 1.0, 0.8, 0.8, 1.6, and 0.5 respectively. Market move carries the heaviest single weight in the engine, and today that matters.
The day opened with 48 races and 551 runners. The odds-band filter cut that to 444; the field-size gate trimmed it further to 344; the lay-of-day collision check removed one more to leave 343 runners entering the scoring stage. From those 343, the engine produced a shortlist of six horses with composite scores ranging from 7.41 down to 6.91.
Queens Bandit leads the shortlist at 7.41/10, driven principally by a maximum market-move score and the strongest speed rating in the group relative to her field. Weleyff sits second at 7.18/10 — the most complete profile on the board, with actual form figures ('3110'), a class drop, and a solid SR 90 — but her top composite score is still 0.23 points adrift. Guitar Star lands third at 7.06/10, also a steamer, but racing in an unrecognised class at Downpatrick with no going data and no form string on file, which limits confidence. Lady Lena posts the sharpest raw speed in the entire shortlist — SR 159, z-score +2.37 — and decent form ('32-1'), but the market showed no movement at all, a flat +0.00, which the highest-weighted feature in the engine translates as a significant drag. God of Fire and Value Pending round out the six at 6.92 and 6.91 respectively, both steamers but scoring beneath the bar set by Queens Bandit and Weleyff. What tipped the balance at the top was the combination of the maximum market-move reading and a speed figure that is materially clear of her field — SR 156 against a mean of 120.8.
The read I keep coming back to is the market-move signal. Queens Bandit registers a +2.00 on the market-move feature — the engine's maximum, indicating a strict majority of bookmakers shortened her overnight with none drifting. At 10/3 (decimal 4.33), the engine puts her estimated win probability at 37.6% against a market-implied 20.6%, a positive edge of +62.8%. That kind of divergence between what the price says and what the model sees is exactly the territory I want to be in. The speed figure backs it up: SR 156 against a field mean of 120.8, a z-score of +0.98, contributing +0.63 to the weighted composite via the 1.5 speed multiplier. She is meaningfully quicker on the clock than what she is lined up against in this Class 4 Wolverhampton maiden.
Charlie Clover's yard adds a further layer. A 27% strike rate over the last 30 days — 4 wins from 15 runners — is properly hot form for a trainer, and at a weight of 0.8 it contributes +0.42 to the composite. That is not a throwaway number. A trainer firing at that rate does not send horses to the track without purpose.
The gaps in the profile are real and I will not paper over them: form is unavailable, class-change context is unavailable, going history is unavailable, and draw data does not yet exist in this version of the engine. Three of the seven features are blank. The composite of 7.41/10 is built on four scored dimensions — market move, speed, and trainer strike rate doing the heavy lifting, with form and going absent. That means this is not an 8+ conviction call. It is a solid-to-strong score, but the blanks are noted.
The obvious risk is that three key features — form, class change, and going — are all unavailable, which means the composite is built on a narrower evidential base than I would ideally want. A maiden filly with no form string on file is an unknown quantity; the speed rating tells us something, but without race experience context it is harder to weight. The Wolverhampton AW surface is specific, and with no going or surface history on record, we cannot say how she handles it. If the market has simply been pushed in by connections without underlying performance evidence, the edge evaporates. Weleyff at 4/1 in the Newbury nursery is the runner I would be most concerned about backing instead — she has a complete profile, a class drop, and a steamer read of her own, and she is only 0.23 points behind in composite. The conviction band here is MEDIUM for good reason.
Queens Bandit · 10/3 · 1 unit · MEDIUM
The Fox’s Wire · Bet of the Day
The algorithm pulled five contenders from today’s card. One earned the ticket — the rest are what it beat. When each race settles, the finishes fill in so you can judge the shortlist, not just the selection.
Position only — payout-neutral. A process check, not a P/L. Click any horse to read its profile.
